Careers, task by task

Family medicine

Care across a whole life: ambiguous symptoms, prevention, chronic disease, and knowing people for years.

Training: College, four years of medical school, then a residency of 3 or more years
Today's median pay, May 2025: $244,180
Projected job growth, 2025 to 2035: 3%

U.S. Bureau of Labor Statistics, Last modified August 27, 2026

Occupational Outlook Handbook: Physicians and Surgeons

Scope: United States.

How it was done: Federal employment and wage statistics. Pay is the May 2025 median, half earn more and half earn less. Growth is a 2025 to 2035 projection.

Limits: A median across all ages, regions and employers, before tax. A projection, not a promise. It tells you nothing about one person's income. The outlook text on this page names aging and chronic illness as demand drivers, and does not mention AI.

Checked against the original on November 6, 2026.

Open the source

A snapshot of people practicing today, not a forecast of your income.

The work, and what we can verify about AI in it

Task lists are our editorial summary of what the role involves. They vary by setting. The status describes the evidence we've checked, not whether a job is safe.

Taking histories and examining patients
Judgment and accountability
Insufficient evidence
Insufficient evidence. We haven't verified a solid source on AI in this task yet. That's different from evidence that nothing is happening. If you know of one, send it to us.
Writing notes and charting
Information work
In real use today
- Established now
  AI scribes saw the broadest adoption of any clinical AI category in 2025, reaching 63% of hospitals using Epic's record system. A prospective Stanford study of 48 physicians found a median 20 minutes saved per half day of clinic.
  What this does not prove: Measured mostly among physicians, so results for family physicians may differ. Hospital systems reported many of these results themselves.
  Stanford HAI, 2026
  The 2026 AI Index Report, Chapter 6: Medicine
  Scope: Global research, with U.S. data on FDA authorizations.
  How it was done: An annual review that summarizes studies by other researchers.
  Limits: A summary of studies with different designs. Each headline figure comes from one setup and does not generalize on its own.
  Checked against the original on November 6, 2026.
  Open the source (opens in a new tab)
Screening for diabetic eye disease
Information work
Narrow use or authorization
- Established now
  FDA authorized IDx-DR in 2018 to automatically detect more than mild diabetic retinopathy in primary care, finding 173 of 198 cases and correctly clearing 556 of 621 people without it. Medicare's 2026 value for this test includes zero physician work units.
  What this does not prove: One disease, one camera, adults with diabetes only. It screens and refers. This is the clearest case of a task moving, not of a job disappearing.
  U.S. Food and Drug Administration, April 11, 2018
  De Novo DEN180001: IDx-DR
  Scope: U.S. adults with diabetes and no prior diagnosis of retinopathy, at 10 primary care sites, 819 fully analyzable.
  How it was done: FDA's review of a pivotal study of automatic detection using one camera model.
  Limits: Screens for one eye disease and refers. It does not treat, and FDA says physicians should review results and advise on referral.
  Checked against the original on November 6, 2026.
  Open the source (opens in a new tab)
Renewing prescriptions
Judgment and accountability
Narrow use or authorization
- Contested
  Utah lets Doctronic's AI process renewals of about 190 existing, non-controlled medicines. In the current phase, a physician approves every one.
  What this does not prove: The state's medical board majority urged suspension, and a move to less review has not been approved. The listed term has ended, and we could not confirm its status since.
  Utah Department of Commerce, Undated page, with a letter from April 21, 2026
  AI regulatory relief pilots, and the Doctronic prescription renewal agreement
  Scope: Utah, for renewals of about 190 existing, non-controlled medicines.
  How it was done: The state's own pilot pages and correspondence.
  Limits: Contested: 11 of 14 Medical Licensing Board members urged suspension. The listed term ended October 24, 2026 and we could not confirm its status since. Utah says it lacks robust evidence of benefit.
  Checked against the original on November 6, 2026.
  Open the source (opens in a new tab)
Drafting replies to patient messages
Information work
In real use today
- Established now
  19% of physicians in the AMA's 2026 survey use AI to draft replies to patient portal messages.
  What this does not prove: Self-reported, across all specialties.
  American Medical Association, March 2026
  2026 Physician Survey on Augmented Intelligence
  Scope: United States. 1,692 physicians, fielded January 15 to February 2, 2026.
  How it was done: Voluntary, self-reported survey. Full and qualified partial responses are included, and each percentage reflects those who answered that question.
  Limits: Reports what physicians say, not what happens to patients. The question list changed between waves, so the AMA itself cautions against strict comparison across years.
  Checked against the original on November 6, 2026.
  Open the source (opens in a new tab)
Managing several chronic conditions over years
Judgment and accountability
Insufficient evidence
Insufficient evidence. We haven't verified a solid source on AI in this task yet. That's different from evidence that nothing is happening. If you know of one, send it to us.
Building trust across a lifetime of care
Human relationship
Insufficient evidence
Insufficient evidence. We haven't verified a solid source on AI in this task yet. That's different from evidence that nothing is happening. If you know of one, send it to us.

Skills that transfer

Clinical reasoning
Communication
Population health
Care coordination

What nobody has measured

Whether AI triage reduces visits or uncovers more need.
How panel sizes and pay change as routine contact automates.

Nearby routes

Compare these side by side

Family medicine

The work, and what we can verify about AI in it

Taking histories and examining patients

Writing notes and charting

Screening for diabetic eye disease

Renewing prescriptions

Drafting replies to patient messages

Managing several chronic conditions over years

Building trust across a lifetime of care

Skills that transfer

What nobody has measured

Nearby routes

Read next